Myeloid Sarcoma Reveals a Unique Tumor Micro-Environment and Differential Prognosis Based on Transplant Status and Type of Treatment Received

Background

Myeloid sarcoma (MS) is the rare extramedullary proliferation of myeloid blasts that disrupts the normal architecture of various tissues in patients with Acute myeloid leukemia (AML). MS can present concurrently with untreated AML or present as relapsed AML. Based on our clinical observations and emerging immune checkpoint inhibitor (ICI) studies in AML (Davids et al NEJM 2016), MS appears to behave differently based on transplant status and tends to have better ICI sensitivity. One of the prominent mechanisms of relapse post-transplant is immune escape via the loss of leukemia recognition by the graft. This is mediated by down regulation of HLA expression. We hypothesize that the immune signatures in the bone marrow (BM) microenvironment are distinct from those from the local MS microenvironment. Thus, we aimed to comprehensively explore the MS microenvironment at single-cell level, providing clues for more effective therapies including ICI, and insights into the molecular pathogenesis, clinicopathological features and outcomes of MS.

Methods

This is an investigator-initiated study where we collected paired fresh MS tissues and BM samples from patients with MS as well as retrospective collection of paired MS and BM samples that are already sectioned and paraffin-embedded. Eligible patients were consentable adults > 18 years old and have MS lesions whether at diagnosis or at relapse of their AML disease. We excluded patients who have multiple malignancies or patients who are unable to provide samples. Accepted samples are 10-millimeter tissue or 1 million cells per milliliter. The fresh samples were cryopreserved and used for single cell RNA sequencing and ex vivo immune cellular assays. The sectioned tissues were used for multiplex immunohistochemical staining (mIHC) that tested a panel of 6 antibodies (CD3, CD68, CD4, CD33, PDL-1 and HLA-DR). We sought to further perform spatial transcriptome analysis (Visium) that complements mIHC for in-depth characterization of the MS microenvironment and immunopathogenesis. The clinical and genomic annotations for these samples were abstracted from the patients' charts. We have utilized Next generation sequencing (NGS) panels that interrogated 24- 180 oncogenes, including FLT3, IDH, CALR, NPM1, KRAS, NRAS, and TP53. Survival estimates using K-M curves and multivariate analysis were performed via SPSS.

Results

Our study recruited 30 patients so far. Mean age at diagnosis was (57.6 ± 14 years); 63.3% of the patients were males (n=19) and 80% of the patients were Caucasian (n=24). In our cohort, 46.6% of the patient had MS diagnosed as the cause of relapse after stem cell transplant (n=14) while the rest had MS at the initial diagnosis of AML. The most common cytogenetic abnormality in our cohort was complex karyotype in 30% of the patients. Molecular landscape of MS at diagnosis and relapse was done in 90% of the patients and shown in figure 1A,B. MS conferred a poor prognosis if diagnosed on initial presentation vs a new presentation at relapse post stem cell transplant with a median OS (4 months vs 25 months p 0.003) (HR 3.9 - 95% CI 1.4-11.1) figure 2. Interestingly, we have noticed that patients who received Venetolcax based therapies have poor prognosis compared to other therapies (IDH/FLT3 inhibitors, conventional chemotherapies, or ICI); median OS (9.3 months vs 24 months p 0.02) (HR 3.9 - 95% CI 1.2-12.6) figure 3. In a prelim analysis using validated mIHC, we have observed abundant HLA-DR expression in both blast cells and nonmalignant stromal cells of the MS microenvironment (n=4) figure-4. By contrast, PD‐L1 was expressed mainly on a number of stromal cells rather than blast cells in the MS microenvironment. Further special localization analysis of tumor immune infiltrates (TILs) is being conducted to disclose clinical outcomes and biological significance of PD-L1, HLA-DR and TILs in MS. Spital transcriptomics are in process and will be presented.

Conclusion

MS diagnosed concurrently with untreated AML appears to confer a poor prognosis and behave differently than MS diagnosed as relapse post-transplant. MS appears to be resistant to Venetoclax based therapies and patients tends to do worse on this therapy. In contrast to BM microenvironment, blast cells within MS appears to have abundant expression of HLA-DR in both neoplastic tumor cells and nonmalignant stromal cells promoting immune recognition which might explain their sensitivities to ICI.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal